Sample Data (not enough of it)
Sample data provided when configuring new P2GL formulas is limited to 20 or 25 rows of transactions, which typically is not enough sample data to ensure that formula configurations are accurate and doing what the client needs. Two examples:
In one case there were transformations configured that looked right on the first 20 rows of data, but we couldn't see that the logic was incorrect on other rows. It was not until the client ran the post and saw an error on row 190 that we were alerted that the logic needed to be refined.
In another case, skip row logic was used to exclude transactions with foreign currency from posting, but the first 20 rows of data were ALL rows with foreign currency, so all sample data rows were skipped and there were no rows of data available to validate configurations.
Bottom line is that we need more than just 20 or 25 rows of sample data for validation when configuring P2GL formulas.
Stu Manewith
shared this idea
-
Stu Manewith
commented
This is still a fundamental issue with P2GL (including POM) and often requires client disappointment and unnecessary rework. And, even rework typically requires 'flying blind' since a larger volume of transactions are not available as sample data. In addition, clients get annoyed by being asked multiple times to refine their source data so that we can get the necessary sample on which to work.
-
Stu Manewith
commented
I am still very much struggling with this issue. I'm working with a Strategic client who processes both online and offline gifts coming over from Salesforce -- but their current focus is processing offline gifts while we work out some kinks with online. The majority of gifts are online, so a skip row function removes them, but also removes all the sample data coming through -- and our P2GL filtering is not yet sophisticated enough to allow me to pull in only offline gifts. I really need some help here.
-
Stu Manewith
commented
This is becoming a more pervasive and challenging issue as we continue to work with more Salesforce customers for Post-to-GL. Because of the need to skip rows within a formula because of inadequate front-end filtering, we don't get ANY sample data at all, and, therefore, need to run the formula and then examine the output, iteratively, in order to test and refine the formula logic.